-
wmmmm. - mmm »8HBB ^miwmm
^mmMm'v^T^»mm'''cw?:m
E 1 [ I I I
I i
UUCLASSIHtL)
HnmRRO
Wi|Ä!sf**-
/V
Date >SVVUiLxcV> i^^
o o
r'-. i ■ -. ii ;■ i j i-lümRRO Cpnirot No
Ä -88;?4 ■M ..:..i
^/^Gi-'-yj?. COPIES
fOR OPEN PUBLICATION
yi OCT 0 Z 1975 6
mj A.,[ if. ... (0AS0_pAJ
':;::,"■ •■..! .- ;.-r(-f(j..-
DOWNGRADED ^ 3 V^R WTEWALS DECLASSIFIED AFTER 12 YEARS
D00 DIR S200.10
The George Washington Uiiiversity HUMAN RESOURCES RESEARCH
OFFICE
operating under contract with THE DEPARTMENT OF THE ARMY
C2059 Approved for public release
distribution unlimited it
"'':'-'-" ^^.i^Ufc^i^d^a^jyaetfa^^ . ^^^ai^^^ «ilrimrtiilDi mmam
mummmm
-
HnMWMtM «MWHI:
iiNi:iAvMH(-n
Tiis material has been prepared for review by appropriate
reaearcn or military agencies, or to record research information on
an interim basis.
The contents do not necessarily reflect the official opinion or
policy of either the Human Resources Research Office or the
Department of the Army.
ii II 0
/
V
ii
r 1 :
* i The Human Resources Research Office is a nongovernmental
agency of The George Washington University, operating under
contract with the Department of the Army. HumRRO's mission is to
conduct research in the fields of training, motivation, and
leadership.
01 n l :
^^^^^^^^^^tlt-Tf^Ht
n
■.«■..•:a-..i«*i«*i>(.t: Ai'-üii*Uik: ■vrMiM,Mk*lM*k ■ ■ - -—
'— ^"——^^^—^
-
mmmmmim-im i$w*
I
■^'■^■-■^•'■■■■^'^•^•- ■' :*Tmr,-)r-*'''* ,.- '■'■"■■^ ^^•,:
■,-r:^w...i
-
mmm mmm* v. K.inmjm
nummMmm murnmam* »w *»»«ww*«(*»¥'«» mrj«.»»e»'S*!'l»»-'»w"
wilwwwil|iiwm»Ml»^iliilM^ 11 lnmuwWWUKMtMIBIHWPtMIÜilWI
I
n
li I n
FOREWORD
This Consulting Report presents the results of tests conducted
by
The Human Resources Research Office, Division No. 5 in support
of Joint
Task Force Two (JTF-2) of the Joint Chiefs of Staff.
JTF-2 was organized by the Joint Chiefs of Staff to conduct a
series
of coordinated and integrated tests to determine the
capabilities and
vulnerabilities of offensive and defensive weapons systems
operating
in the low altitude regime. Test 3-1/3-5 Nonfiring (NF) provided
discrete
operational and tactical data for the evlauation of visually
sighted and
radar controlled air defense weapons systems against low-flying,
high
speed tactical and strategical aircraft.
This report is concerned with the results of that portion of
Test
3.1/3-5 (NF) which was conducted in the Oklahoma/Arkansas
environment
of Test U.l, Visual Target Acquisition.
i
The data reduction and statement analysis for this test was
provided
to JTF-2 by Mr. Michael Carter of Sandia Laboratories,
Albuquerque, New
Mexico. The support of the Human Resources Research Office
(HumRRO) was
provided under authority of the Department of the Army in
accordance with
the HumRRO prime contract, DA W-188-AR0-2.
0 D
ROBERT D. BALDWIN Director of Research
iii
PBECEDim PAGE BUMUNQT fllMSD
" ^■^•*-^!^:.»^ .-*a^-^WA.^.^ fe^a**! "*-**** ■*•*"'L*
-
"^mm^^^mmmmimmmmimmm PII»WP«BHIIHI»I mm
i
*Wffß^tf.!tlllj(0Mmrm ü
W..l'J-flf«M^1t^,'>lTlW^V"A.ri.mJM
f
t:
a ft
I I i» ■
I I I I
waA-w^ it
SUMAFX
1. (U) Joint Task Force Two Teat 3.1/3.5* Surface Based Air
Defense (RP),
was designed as a series of coordinated test efforts to provide
opera-
tional and technical data for the evaluation of the capabilities
of surface
based air defense systems against attacks by low-altitude high
performance
aircraft.
2. (U) The Oklahoma/Arkansas portion of Test 3.1/3.5 (KF)
investigated the
capabilities of ground observers to detect and estimate the
range to air-
craft flying attack missions against prebrlefed ground targets.
Observer
performance was obtain ad for three aircraft flying three
programmed speeds
and two programmed altitudes.
3. (U) The trz'j environment used was the sane as that for Test
^.1, Visual
Target He-ignition. The Test 3-1/3.5 (HF) portion of the test
was conducted
by Human Resources Research Office (HumREO), Dlvison No. 5 (Air
Defense) on
a cooperative noninterference basis. Instrumentation and data on
event
times as well as data on aircraft position, speed, and altitude
were pro-
vided from that collected for Test k.l.
h. (U) During ihe conduct of the test, considerable aircraft
position
data were lost due to faulty Instrumentation and offcourse
aircraft flight
paths. During the data reduction portion, additional data losses
accrued
due to ambiguities in observer responses and missing visibility
and unmask
data. As a result, the original test objectives for Test 3.1/3.5
(RF) had
to be considerably modified. Final objectives were to
detemlne:
a. the relationship between the frequency of detection and
the
slant range to the aircraft.
I
mAaia^iiü^äiääiäaiäis^ij^ «II Mfc PRSCSDINJ PAGE BUHK-NOT FIUQD
Maiaiaaaaiiigii^tüiiiü MM
-
•mv^yw ■W^MMMQin
\M o\ ^
-
W^f^mmmmmm |II^,I.,>IIunniwi i„ i HI ,1 imypm it,,,iipi,
.IIMHHH^^,,,,....!.V^WWT^. fmniftjH w*
KtWWMMMMWIMani —MM—MIMIWWBN<
:
i
I
I
I
i
I H
I I
■I
^
I I I I I I I I I I I I I I I I E I
H
6. (U) The teat data were recorded on magnetic tape by the G8IP
located
at each target. These data were eub8eq.uently edited to remove
ambiguities
and time related to Test 4.1 position data. The resulting
responses were
edited further, and any data not considered plausible were
removed. Further
data reduction consisted of preparation of analysis tapes ana
the providing
of descriptive statistics. The analysis consisted of a review of
the des-
criptive statistics (histograms and scatter diagrams) for the
purpose of
formulating initial analysis of variance (ANOVA) and analysis of
covariance
(ANCOVA) models. These models were processed using the weighted
regression .
analysis program (WRAP) and multiple analysis of variance
(MANOVA) statisti-
cal analysis computer programs. The resulting series of analyses
revealed
that the test variables were strongly confounded. This
confounding and
the unbalance in the design largely precluded any meaningful
results other
than from descriptive statistics. A further series of
examination revealed
the individual flight profiles as the source of the
confounding.
7. (C) The average distance of the aircraft at detection for all
trials
was approximately 6,200 meters. A cumulative percent detection
curve as
a function of slant range was obtained for all trials, and these
data were
compared with data from two previous detection studies.
Considering that
the three tests were not comparable with respect to aircraft,
test environ-
ment, or instructions to observers, the range of differences in
detection
ranges for the three studies were not considered to be
unusual.
8. (C) Cumulative percent detection as a function of slant range
was
obtained for each of the four ground targets separately. The
average
slant range at detection varied over targete between 5^300 and
7*700 meters.
Differences between the targets were partially explainable in
terras of
differences in gross terrain and unmask characteristics of the
four targets.
9. (C) A measure of aircraft apparent size (ASA) was found to be
correlated
with cumulative percent detection. The regression equation
describing this
,\Nr\ ^s\^ vil
-
aä^ima^i^i^totiiaj^
-
iWPPWPBWWP^I^WPIjWW
Wi^i|PipiluyiWfiiiii)piiMpiiiii|i|ii,i,i:ifi.ikw;i|»ijiwi.M.i''.i'i.ifi.,....wir.'.i
immr-wwyr^w-,:-,,— -
I i.
! i
\
\v Ar\ ^
-
I
TABLE OF CONTENTS
Page
I
II
0
0
Ü
J
n
Li
11 H
SUMMARY v
SECTION 1. INTRODUCTION 1-1
1. BA.CKGROUND 1-1
2. PREVIOUS TESTS 1-3
3. TEST OBJECTIVES 1-5
h. TREATMENT OF TEST OBJECTIVES 1-6
SECTION 2. CONDUCT OF THE TEST 2-1
1. FIELD TEST OPERATIONS 2-1
2. OBSERVERS 2-1
3. TEST SITES 2-2
h. GROUND SITE INSTRUMENTATION PACKAGE (GSIP) 2-3
a. General Description ■ 2-3
b. Unmask Time . . . .' 2-3
c. Aircraft Position 2-3
d. Timing 2-5
e. Visibility and Illumination 2-5
f. Observer Response Boxes (ORB) 2-5
5. DAILY SEQUENCE OF TEST PROCEDURES 2-7
a. Rotation of Targets 2-7
b. Test Monitors 2-8
c. Procedures Prior to the First Trial . , 2-8
d. Procedure During the No Early Warning Trials 2-9
e. Procedure for Complete Early Warning Trials 2-9
SECTION 3- DATA PROCESSING AND ANALYSIS 3-1
1. GENERAL 3-1
.2. QUALITY CONTROL OPERATIONS 3-1
3. DATA REDUCTION 3-2
k. ANALYSIS PLAN 3-2
5. EFFECT OF DATA REDUCTION ON THE ANALYSIS PLAN 3-3
I ha* .vim mmm i ■ mmnm
'titi&äjk&iC ^■^^^ii^,Li'tv:.ü,u^JiÄ^_^.,;ÄJ....^i^^«J(V
^i^j^*ri^jii^wi.^w.ÄÄt^^i^ait£iiädii^ iiiiMiiiitiaiiiiiiliiBii^
-
IfJHp^WJIWWIMIJP'lWIIlW" WPiNi. ) Hii.iiH ^pnq
l1!f.».l|J|,ilii|;p»»i^>Wiyiy,lij1il^,jf|i.,,jl,T?niTT^F,OTt!^r
TABLE OF CONTENTS (continued) Page
SECTION k. RESULTS AND DISCUSSION k-l
1. INTRODUCTION ^-1
1'. DETEGTIO:; ^-1
j. FOSITRIAL QUESTIONS ^-T
a. Scope ^"7
b. Questionnaire Results ^-7
c. Aircraft Background ^-7
d. Smoke k-J
e. Visual Versus Auditory Detection k-9
h. Tlf-E BETWEEN PILOT ACQUISITION AND OBSERVER DETECTION ...
k-9
a. Introduction ^-9
b. Performance Over All Trials ^-10
c. Differences Among Targets ^-10
d. Trial Effects . . . ' ^-13
e. Discussion 1+-15
5. POSTTEST INTERVIEWS ^-15
6. RANGE ESTIMATION ACCURACY k-l6
APPENDICES
A BIBLIOGRAPHY A-l
B RANGE ESTIMATION TRAINING B-l
C ARKANSAS POSTTEST INTERVIEWS C-l
D DATA PROCESSING AND ANALYSIS D-l
FIGURES
?.-\ Observer Positions; Target E3 2-k
k-1 Percent Detection Compared with Prior Field Studies ....
k-2
k~k Cumulative Percent Detection by Ground Target Site ....
^5
h-'Jj Computed Unmask Profiles for Each Ground Target ^-6
k-k Distribution of J k-11
J \
. J
y
Mii«WWWfillIMW»lt(ll»™w''^',~
iiüüi
-
I^P^IIflH^W^^
I?. IWWWMTflVMlWi'Pir^^^ «imwiWCM illWM*i-fi'itM»W
-
P^AWflWIBBIW IWBiPPPI WIJI|IW||l|iWi|ppi
-
^BUmSSBm BB -^ i-;. R "SSWraR! ™- ÜUPPPWIBWIBPlii
Mi,niiiiiin,..n«»uimi»,u
e. JTF-2 Test 4.1 was an air-to-ground visual target acquisition
test
conducted in the Oklahoma/Arkansas area in the general vicinity
of Mena^
Arkansas. One of the objectives of Test h.l was to deteimine the
abilities
of representative aircrews to visually acquire prebriefed ground
installa-
ticns consisting of simulated military targets such as
surface-to-air
missile and radar sites, landing strips, logistics supply
points, and
bridges. It was concluded that this air-to-ground target
acquisition
test would provide an environment for obtaining ground-to-air
observations,
such as visual detection and distance estimation judgments.
f. Test 4.1 was conducted under conditions that offered a
unique
environment for obtaining ground-to-air visual observation data
and would
supplement the very large amount of observer data obtained at
HDffl.
U
(I I i
;
(l) Although the aircrews participating in Test 4.1 received
normal preraission briefings concerning the location of the
ground targets,
each aircrew would participate in only one flight over each
target.
Consequently, it was expected that the crews would exhibit
variability in
their navigational and target acquisition ability. This
variability in
the performance of the aircrews would increase the uncertainty
of the
ground observers concerning the time of arrival at the target
and the
direction of "attack" by the aircraft. These conditions of
uncertainty
concerning the aircrafts1 time of arrival and direction would be
more
representative of tactical ground-to-air defense than if the
aircraft
always flew the same flight path to each target.
(2) One measure of aircrew performance of interest to Test
4.1
was the time and distance at which the aircrew visually acquired
the ground
target. Since this pilot acquire event was recorded for each
trial during
Test 4.1, this event would be compared with the ground observer
detection
time and distance events to obtain data relevant to the
following questions:
(a) Does the air defense gunner tend to detect the attacking
aircraft before the aircraft locates the defended point?
1-2
i i
..-._J,L, .:-, ,^ .. ■-. . ^tr.^oi,.«.; ■,
-
PPaoWCTW^WwaawwipwwaHaainpni . WWM«
t
: I
WWIIIMBWIIWIWBMWWWW^W^
^aiW^iyWWppWpWWfH|ffWfljl»fl-l>u|.i-i||i.W.iW
I
11
0 Ü
n Li If r i
.; I.]
n
f i 4 i -I II
1 I
0
(b) When the pilot ac^uires the target first, how much time
elapses before the air defense gunner detects the aircraft?
(3) The instrumentation provided for Test k.l included
distance
measuriiig equipment (DME) for accurately locating the position
of the
aircraft; a radar line-of-sight system between aircraft and each
ground
target for determining the time at which the unmask and remask
events
occurred, and visibility recording equipment at each target
site. The
DME and electronic unmask instrumentation provided accurate data
concern-
ing the slant distance to the aircraft from the target when the
visual
detections occurred and permitted computation of
unmask-to-detection time
delays. Measurement (computation) of the unmask-to-detection
time delay
was a potentially important performance measures that could be
used to
evaluate the effects on visual detection of controlled
(independent) test
factors, such as variation in predetection altitude, aircraft
speed, and
the accuracy of the early warning provided to the ground
observers.
2. PREVIOUS TESTS.
a. Visual Detection. A number of tests have been conducted in
desert
terrain concerning human ability to detect low-flying aircraft
under various
simulated tactical conditions.
(l) Tests conducted at Gila Bend, Arizona, by the US Army
Human
Engineering Laboratories (see reference 1) varied the size of
the sector
to be searched, but provided no temporal early warning (no
information con-
cerning probable time-on-target and time of arrival). When the
search
sector was '+5 degrees, the mean distance of jet aircraft (T-33;
F-86, and
F-100) at detection was 2,750 meters. When the search sector was
90 degrees
and 360 degrees, the average distances were approximately 2,585
and 1,985
meters, respectively.
1-3
1 ww»**«-,**«*»»**««:^»*«.-* MmiWiMiMiiwtK'wiimMMawiMaiaBM^
iaäitämääimä^ääiMäiMiäMKiMäiäiiiilMä iMiiiiigiiii
-
JRHMP^UfV^PRlTP i- \'.~^'-
■**W:- *■'■ ■Kt^^KW.w^.^WKVt'-t'
(2) Tests with T-33 and F-100 aircraft conducted by White
Sands
Missile Range (WSMR), New Mexico (see reference 2), used
30-degree and l80-
degree sectors and varied temporal early warning from lesa than
15 minutes
to more than 90 minutes. The mean detection range of F-100
aircraft, for
example, averaged over all search and early warning conditions
was approxi-
mately 4,1*00 yards. However, detection range appeared to be
influenced by
aircraft heading angle, altitude, and speed a« well as other
factors. The
affect of the lengthy alert intervals on detection range could
not be
evaluated.
(3) Tests were conducted by The Human Resources Research
Office
(HumRRO) in April 1965, using the same terrain used earlier for
the WSMR
tests. The HumRRO tests ccmpared unaided versus binocular-aided
detections
using a 30-degree search sector and 1 to 5 minutes of early
warning. Under
these conditions, the average range at which jet aircraft
(F-l+C, F-100, and
T-33) were detected was approximately 10,000 meters, when
averaged over all
viewing systems and observer offsets. The Increase in detection
range,
as compared to earlier tests, was attributed to the increased
accuracy of
the early warning information provided the observers concerning
the heading
and expected r.iie of arrival of the aircraft. When averaged
over all air-
craft, the optical aids did not reliably increase detection
range.
{k) Tests were also conducted by HumRRO in June 19^5, in
conjunction
with the JTF-2 Test 1.0, Minimum Terrain Clearance, at Tonopah,
Nevada.
Either 1 or 5 minutes of early warning was used, and the exact
heading of
the aircraft at the time of unmask, was known. Both near and far
terrain
masking existed, and unaided and binocular-aided detection data
were
obtained. The mean detection range for F-^C and F-105D aircraft,
averaged
over all viewing conditions for the far terrain masking
condition, exceeded
12,000 meters. Detection range was not increased by using either
6 x 30 or
7 x 50 binoculars, nor did the difference between 1 versus 5
minutes of
temporal early warning have a reliable influence on detection
range.
1-4
waikki'iJitÄfc^dk&s^ -K"-'"t;"i;"'iiii[ttritBiii'ii
-
i;»!!^;^*»^«.^ WRrmn? mnmwi fn "■rfw*,"'■
..■.flf,^'-' ■" »»"S/WWTtff«"* ViWi1," ■■ l>1(..i*1i^!"(*rff
.-.:r-V),',v,T^..*',i:-,'.'V;'
■■^^-,^^^^v■•.,■!.■^-.t.^f'.^A^J•^frl''K^■.^•^^Wf>tkK^^■^■'^?^'■■■l•^'V,•>^*«■-
l5Ml^>«wl%WP^^'V■r«h^^■ütl^(r^.;>A'v*7^*'f«r.«■«*,W
1 Il
0 f 1(1 I in 11'
(5) In summary, the previous visual detection tests had been
cai-
ducted under good visibility conditions in desert terrain. Most
tests have
occurred under far terrain masking conditions. The studies
collectively
indicated that increasing the precision of early warning
information
markedly increased detection range, and that detection range was
influenced
by aircraft altitude, aspect angle, size, and speed. However,
the extent
to which detection ranges experienced in the desert can be
generalized to
other terrain and meteorological conditions is not known. In
addition,
tests of the effect of aircraft size (type), speed, and near
terrain
masking on detection were needed.
b. Range Estimation. Very little test information exists
concerning
nan's ability to judge when high speed aircraft are within the
effective
range of air defense weapons. In I966,, HumHRO ccnducted several
studies
comparing different methods of training gunners to estimate
engagement
distances of 400, 800, 1,500, and 2,500 meters, the approximate
open and
cease fire ranges of various forward area air defense weapons.
Although
the training did reduce gross judgmental errors and reduced the
variability
of judgments among observers, all training and testing had been
done in
desert terrain. As a result, the affect 'n ground-to-air
distance Judgment
of different aircraft backgrounds, different types of terrain,
and aircraft
size were not known. It was also not knownhow well the
estimation skills
developed during the desert training would transfer to other
type of terrain
and aircraft.
5 ■
Ü 3. TEST OBJECTIVES.
1 n
I
a. As originally conceived, the tests of visual detection and
rarge
estimation to be conducted in the Oklahoma/Arkansas portion of
Test 3.l/
3.5 (NF) were designed to evaluate the effects of specific
independent test
factors, such as the scheduled variation in aircraft altitudes,
speed and
type, the accuracy of the early warning (EW) information which
would be provided
1-5
1
W»u»»>*/.»>-^i'*Vi[Jw*.j^^.«v.~vw-«««i
-
r^T^w^W'iv^'^Z^^fi^ W-iMl,4|iPWI|HmWHipBI)M|»jfmy;t;.i1
•1 < '1
the observers, the time and distance at which the aircraft was
unmasked
from the near terrain or horizon, and the ambient visibility and
illumina-
tion conditions.
b. The tests that were designed to evaluate these factors of
interest
to Test 3.l/3.5 (NF) were based on the following assumptions
conceming the
results of Test 4.1.
(1) The time and distance at which each aircraft flight
became
unmasked to each ground target would be available for use in
ccmputing the
fundamental unmask-to-detectlon performance measures and for use
as a
covariable in the evaluation of the effects of the test factors
on the
performance measures.
(2) Valid visibility measurements would be provided for each
trial
(defined as the flight of an aircraft over the target area).
Variation In
the visibility measurements from trial-to-trial would be used as
a covariable
in the evaluation of other test factors. Visibility variation
also would
be examined for Its effect on the visual observations.
(3) The variation that would exist among the aircrews in
their
ability to navigate to each ground target would be equated for
each com-
bination of test variables, such as the three aircraft speeds or
the two
programmed flight altitudes. This assumption was critical
because of
technical requirements of the planned statistical analysis.
$ k. TREATMENT OF TEST OBJECTIVES.
a. The original objectives established for the
Oklahoma/Arkansas
portion of Test 3.1/3.5 (NP) were not achieved due to problems
associated
with the instrumen+atlon used for measuring the aircraft unmask
and visi-
bility conditions, loss of ground observer event data, and large
variations
1-6
^.s^Mi^feikJb^a»^^
-
W9E^iWK^^nMB^ssMiS!amv^mmmai!w^m!»' -~,~-. •- ^^P ^WWWWmiU'fWW
WW» H'Pwwwyw
\-»vtwKH™*^mm""m-'''Wty: -•''w™-.ii^'t'»*m>PMmtm«it
e*miwmsii'vm*vwm •«•««»».1 ^-•twir*.™^..»«^'
(-■.M',"'JV,i>''«*''*wf^
B
i 0
n
in the altitudes and flight paths of the aircraft. The extent
and magni-
tude of these problems was not ascertained until after the data
collection
portion had been completed.
b. During the data reduction portion of the Test 4.1, it vas
found that
there were a considerable number of unaccountable irregularities
in the
unmask time and distance events which were based on radar
sensings between
the aircraft and the ground target. As a result, it was
concluded that the
radar unmask data could not be used in evaluating the results of
either
Test h.l or Test 3»l/3'5 (Nl?) because the time at which the
aircraft became
unmasked at the target was a significant factor in the design of
the air-to-
ground target acquisition tests, terrain surveys were made and
topographical
map studies were conducted after data collection was completed.
These
surveys resulted in the preparation of a mask profile for each
trial which
displayed the aircraft's altitude in relation to the computed
masking due to
horizon and local masking features. Because the exact range at
which an
aircraft became unmasked for each trial could not be determined
with any
reasonable assurance from the profiles, the computed unmask
ranges asso-
ciated with each ground target were used qualitatively in
evaluating the
field test results presented in this report.
c. The accuracy of the visibility measurements made during each
trial
was similarly questioned by Test h.l. It was concluded that
these measures
would not be used in any precise or quantitative analysis of the
field test
results reported here.
d. The nonavailability of reliable measurement of the unmask
events
and visibility conditions made it necessary to limit the test
objectives.
After extensive inspection of the test event data and the
aircraft flight
profiles, it was concluded that valid data was available for
evaluating
the following test objectives.
in
1-7
imtmm'iMii&iHßi'täVt'V.-r^-mi&v-.-, -■^^vmen
■ ,^.w^.r, . WLvJa^^iMU^vtlii^^
-
itWiWW^^jj^^
SBW^BMaww i imiumKmmimmrifm .^^»v«*»»»»,, ■..
(1) Determine the relationship between the frequency of
visual
detection and the slant range of the aircraft summed over all
observing
conditions.
(2) Determine the relationship between the frequency of
visual
detection and the apparent size of the aircraft summed over all
observing
conditions.
(3) Determine the relative frequency viith which the aircrew
visually acquired the ground target before the aircraft was
detected by
the observer.
(4) Detemine the relative frequency with which the ground
observer detected the aircraft before the aircrew visually
acquired the
ground target.
(5) Describe the relationship between the frequency of
visual
detection and the distance to the aircraft for each ground
target
separately.
(6) Determine the accuracy with which ground observers judge
engagement distances of 400, 800, 1,500, and 2,500 meters under
conditions
different from those used during training.
. 1 1
! I ■ 1
, f t 1
ü i
I ill
1 m
1-8
a
-
' •"r** WSRTBBC ip^IPBp^^ i tgpg ^iWtJipP|pBppwppip
.-v,vw«' M^«>WW!^W^«Ww«WMW^^
; ii
I 11 ! fl ( li 1
{ )
? i
in n
i
11
SECTION 2
CO:;EUCT OF THE TEST
1. FIELD TEST OPERATIONS.
a. The field test was conducted in a rectangular section of
Oklahoma/
Arkansas approximately 60 KM by ^0 M centered on Mena, Arkansas.
The test
area consisted of rolling terrain that provided natural surface
undula-
tions suitable for aircraft in flight to mask themselves from
visual and
electromagnetic line-of-sight and also to provide visual masking
of certain
target locations. It was made up of forests, meadows, rural
roads, and
cultural feat-ores representative of small to medium population
communities
in the temperate zone worldwide,
b. The flight courses to the targets were arranged to cross
ridge
lines nearly perpendicular to the final reference point (FRP) to
target
line to provide definite unmask events for each target. Each
aircrew was
oriented to the geographical location of the FRP for each target
and had
the task of visually acquiring each prebriefed target.
c. The aircraft used for the 3»l/3'5 portion of the test
included the
F-kOj A-6, and F-105D. These aircraft flew over the target area
at the
speeds and altitudes shown in Table 2-1. A more detailed
description of the
test area and aircraft courses is presented in Report JTP2-4.1,
Volume 2,
"Low Altitude Test h.l, Visual Target Acquisition, Field Test
Description,"
dated October 1967.
1 r* 2. OBSERVERS. Sixteen male enlisted personnel, eight US
Army and eight US Marine Corps, were assigned as observers. The
Arny personnel were
selected by their commanders and came from units stationed at
Fort Polk,
I :.„. 1^ .-. ^ ^Htt, ttMMMmnam*Uimmm^^^m
-
iipwwwaappwiiwpiwip m-fts'wwf.'wrnr- •irt"n,~
Fort Sill, and Fort Hood. vl-.e Marines were selected by a team
composed
of one JTF-2 representative and one Human Resources Research
Oifice (HumRRO)
representative. The Marines were from the 3d- Light AA Missile
Battalion,
Cherry Point, North Carolina. All observers had 20/20
vision,'corrected if
necessary, and ranged in age from l8 to 2k years. The average
General
Testing (GT) aptitude score was 111 and ranged from 83 to
130.
a. Observer Training. The 16 observers, plus noneomissioned
officers
(NGO), were trained in the operation of test instrumentation
equipment and
the estimation of range to aircraft. This training was provided
by HumRRO
personnel at Fort Bliss, Texas, approximately one week before
the field
tests began. An F-100 was used as the target aircraft for the
range estima-
tion training. The aircraft flew 36 passes each of two days.
Equal numbers
of passes were flown at 250 and 750 feet altitude, and the
passes were
equally divided between overhead flights and 200 meters offset.
For one-
half of the passes, the aircraft's heading was north; for the
remainder, the
heading was south. The objective of the training was to have
each observer
accurately estimate when the aircraft was at hOO, 800, 1500, and
2500 meters
from the observer's position on both the inbound and cutbound
portions of
the pass. A more detailed description of the training method is
contained
in Appendix B.
3. TEST SITES.
a. The observer groups were located at four of the Test ^.1
ground
targets. The targets used were identified as West 1 (Wl), West k
(w4).
East 3 (E3), and East k (E4), located respectively at Cherry
Hill, Arkansas;
Plunketsville, Oklahoma; Gravelly and Washita Bridge, Arkansas.
Aerial
photographs of the targets are contained in Report JTF2-4.1,
Volume 2, Field
Test Description, dated October 1967.
b. These targets were selected prior to the test by a team
comprised of
Sandia Laboratory and HumRRO personnel to provide unmask
distances to ground - V 1 ll i
2-2
"" " ' . ■■■..:■■,.,. .„..^ . ■--.-..■. .
|||nn||r't|||fr|rt|||J_ij_^_1H mi^gjiji
-
^'^'"'^'fl^ff'WWywWp^Wgiwiwii; ff i^i> 'im
ii^wffwpwBPwipffliitiwipw-""
l«#^.^>lWr«^^Hl^-Mf»;M^■WW■■*»^nw'-^^■-*^1«wlWJV-JJM«^W«'■.»^^*^
v,WIMTHi«^^^.W»l*n'9WIW^IW,«>W. f ^WMW^^
0 0 0 0 D 0
targets from 3 to 15 miles. (See Appendix C for the observer
locations at
each target.)
4. GROUITO SITE mSTHüMSKTASION PACKAG2 (GSIP).
a. General Description.
(l) Each test site was equipped with a GSIP provided in support
of
JTF-2 Test 4.1. The GSIP telemetered the following information
to recorders
in orbiting C-130 aircraft: Unmask and remask time of the test
aircraft,
the visibility and illumination measurements, and the occurrence
of 4
real time events and 10 nonreal time questions from each of the
4 observer
response boxes (ORB) at each site. A complete description of the
instrumenta-
tion is contained in JT72-4.1, Volume 2, Field Test
Description.
li
(2) Since the distance measuring equipment (DME) had a
predicted
measurement of — 50 feet, the ground observers were located at
specified
points within a circular area having a 100-foot diameter,
centered at the
GSIP box. The actual observer locations depended upon site
characteristics.
Figure 2-1 is an example showing the arrangement of observer
positions for
Target £3- Each observer position was marked with a stake; and
the obser-
vers were geographically oriented to a l80-degree search sector
containing
the expected flight paths of the aircraft. All observers at a
test site
were oriented to the same iSO-degree search sector.
b. Unmask Time. Unmask and remask time of the aircraft was to
be
measured by means of L-band continuous wave transmission from
the test
aircraft to a receiver at the GSIP.
c. Aircraft Position. The location of the aircraft with
reference
to the ground at any instant was detemined by means of DME
carried on
board orbiting C-130 instrumentation aircraft. The C-13O3
received DME
slant range data from the test aircraft. This data, when
combined with
2-3
. , — . .i-^. -ijB- — --■-^■-■-■. ^
-
PffB [I llllill ■piinipi|Wflw»|.. ff uv u ,11 .11,!
vmw.t.t^v,»««N«j-vi
..I.IJP.L.I. WW^TTf ■'.gi1,,!1',"!)- IWv 7m
0 0 |Garp|
© 0
I I*
Figure 2-1 Observer Positions: Target E3
2-1+
L) ;
U
*^&ihk*lZ*ii****itÜi*M ■^'■m^mMmmAm^mmmm^
-
rsnma 1'III,MH;IW.I *^^^mm www
W?»*WlfW«W8W'~vSv.;"/*v#Mr^w ;
'^1*«r»*(!t.>jW'J;-^^,\Vt,"l''r*V"-f^
I I
I n 1 U
: ' ;
D
in i u i
ä n \ Li
i In
111 0
IE
the Di'2 infcrs^ition concerning the position of the C-130
aircraft, per-
mitted coraputation of the ground location and altitude of the
test
aircraft.
d. Timing. Ail event timing vas recorded in inter-range
instrumenta-
tion group (IRIG-B) format.
e. Visibility and Illumination. For Test k.l, each site was
equipped
with instrumentation to measure the sky/ground ratio, total
illumination,
and atmospheric transmissivity (scattering). This equipment
consisted of
an illuminometer for measuring sky illumination and shadow,
photometers
for measuring the sky/ground illumination ratio in the direction
of flight,
and a telephotometer for measuring atmospheric scattering.
f. Observer Response Boxes (ORB), Four ORB were connected by
cable to
the GSIP at each test site. The ORB consisted of two units; a
real time
event unit worn by the observer and a nonreal time question box
placed on
the ground at each observer position.
(l) The real time section of the ORB was operated in the
following
manner:
(a) At the time the observer saw the aircraft, he depressed
the "Detect" button located at the top of the ORB.
(b) When the observer believed the aircraft was at the
speci-
fied incoming range, he depressed the middle, or "inrange,"
button.
i ■ (c) When the observer believed the aircraft was at the
speci-
fied outgoing distance, he depressed the "Outrange" button.
2-5
wawaiiwWMWiiawiwwM mummmimm
M^tiääimm^miMmä^mä^MäiimiMlM^i^iM "^^ -— a -'- mi
ma
-
!!gJg!!i!Si^^ WPS m WIWPPWWIPIPBWI Hiippffwpin imppR nwnw
(u) A shielded spring-loaded error switch was located on
the side of the real liae section of the 0J2. This switch was
used by
the observer to record any accidental batten pushes.
(2) The nonreal tiroe section of the CHB permitted entering
the
following coded infon-iati&n:
(a) observer identification^
(b) incoming range specified for each observer each day,
(c) outgoing range specified for each observer each day,
(d) early warning condition for each flyover (or trial)
(Yes or No),
% \
' i I ■ -•■!
(e) aircraft exhaust smoke was noticed at the time of detec-
tion (Yes or No),
(f) background of aircraft at time of detection (clear sky,
clouds, or terrain), and
(g) aircraft was heard before it was seen (Yes or No).
5- DAIIY SEQUENCE OF TEST PROCEDURES.
a. Rotation of Targets. The l6 men were divided at random into
k
observer groups. A, B, C, and D, containing 2 men from each
service. The
composition of the groups remained constant throughout the
duration of the
tea- to facilitate cycling observer groups to test sites in the
event of
aircraft aborts, inclement weather, or other reasons. The groups
were sys-
tematically rotated to all k test sites during the first 4 test
days. During
the remaining 15 test days, the groups were selectively assigned
to specific
sites to provide maximum utilization of the Test 4.1
flights.
2-6
i i
■r■|TJw■l^ll^^i[mlllllir-^MylliiM|■Tii^■^''"i^^■,^^^*"'" «HUM
m*m mmmmi
-
;.^^Z«*^W^'«**^^,.!«»'''^CI^
v v.-'.■■'■■•".f-'.'"" 'VM'r '.■>M";T.-■'.-"(..,'.'■■^M-
.»(«/»*««i^nani'tf«1««
il ffll*''TR'?B|''*T'l (^ft-,rf'«'/'i¥;j'4*/*''';^i'*(rm'.
«MVHWew-wf.MW.sWWWI ■-*.*.iiw.- ■ ■■ '«^«i1 '^.i:»!*'''."*"»'
f":^*1
|BPSEWB5P8BWff!^iSS
rt.wv.l,^»..;(-,h,v.-:/.w,.«c:-Mji-.
T."wmi''^">v-;,".MnwVi')l-M*
I
1 n I „ I li
I \ u
b. Tost Monitors. Thore uaa one riOiuRBO monitor located at each
test
site. Monitors were rotated periodically between sites and had
the follow-
ing duties:
(1) they supervised the novemeat of observers to the test
sites,
(2) monitors gave observers the instructions for the day and
the
ranges that they were to estiaate,
(3) monitors checked each CHS to insure proper connection with
the
GSIP box and to insure the correct observer code had been put
into the ORB,
»
0 I li
i li
Ml
SB I I f n i li
{k) the monitor, stationed near the communication equipment,
monitored all messages from Test K.l Control and relayed
appropriate
infennation to the observers, such as early warning time, and
aircraft
aborts, and
(5) the monitors recorded the time, course, test site number
and
aircraft type for each trial during a test day. All deviations
or changes
in the prescribed test procedure were recorded, as well as known
reasons
for missing data.
c. Procedures Prior to the First Trial. Immediately upon
arriving at
the test site, the group monitor called Test Control and
reported that his
group was at their site. The group monitor gave the test
instructions to
each observer, which included the incoming and outgoing
distances that were
to be estimated by each observer. Each observer at each site had
1 of l6
combinations of incoming and outgoing distances each day. The
same distances
were estimated for all trials during the day for which complete
early warn-
■ ing was given by Test 3.1/3-5 (N^1) Control. Except for the no
early warning trials, the group monitor announced the time the
first trial was to begin.
Each observer moved to his position and became familiar with the
sector
he was to search on the first trial.
2-7
..fc..i*.wu.,/..,..lei^dt
-
™*BÄIWK (l"^,i»Wl l,,, ,r- ■ U^ ^i'.W'.WDT.'iin'C"-"
d. Procedure ^.--.r.^ ohe Äo Harl/ -/;ü.-v.iri5 "rialö. T'he
observers were
instructed that on certain -crisis thfcy would not be infon^ed
of the exact
tira v>-i-- *i0 c*_rcr'u,J.t Wou^-C a^c^ar [hey «ere
instructed to search a
sector of '+J0 degrees. Anytii.'.e the observer initially
detected the aircraft
visually^ he was instructed to ir:-T.ediatcly depress the
"Detect" button,
even if the aircraft had passed his site. As soon as the
aircraft had been
visually detected, the observer answered the posttrial
questions. There
was no range estimation to be accccpllshed on the no early
warning trials.
1 I
e. Procedure for Complete Early Warning Trials.
(1) When complete early warning was provided, the observer
knew
within approximately — 20 degrees the expected approach of the
aircraft
and the approximate time of arrival, accurate within 1 to 2
minutes. The
observers soon learned that most aircraft approached each test
site from
appro*imately the same azimuth. In fact, sane observers reported
that they
knew the direction after they saw one trial over each test
site.
(2) Test 4.1 Control transmitted information to the test
sites
concerning the aircraft's time of arrival at designated check
points.
This information was used by each monitor to determine the
probable time
of arrival of the aircraft at the test site. This early warning
informa-
tion was relayed to the observers.
(3) On all trials, when complete early warning was given,
the
observers also made two estimates of range, one incoming and one
outgoing.
After the aircraft had been detected, the observers continued to
watch
the aircraft. When the observer believed the incoming aircraft
was at
the assigned range from him, he was to depress the "Inrange"
button on his
response box. He was to continue to watch the aircraft as it
passed over
his position and as it was outbound. When he believed the
outgoing aircraft
was at the assigned range, he was to depress the "Outrange"
button on his
response box. The trial was over when he made the outgoing range
estima-
tion. The observers then answered the posttrial questions.
2-8
mkm^''^^'ie^^iät^s^'m'r'
i.r*:iAvJ.l^tJ-.^^-t-^JlI.-«-vlGifLiEalk^VL^:':.W^.V.:
-
PEPPHBI
i.WfÄ'it". ' iWufcfWWWHBi j^lIHBMMl&WiWMWeaBPMM!^^
Wf^ü^fif^tmmKWwiww'w^^ ■-
ffl«^ «^>*^^y^ft^(f.fm..w^^w-j^*^ • W"' ■3^ ^' ■
immw^ßW^y^^1^
o
0
D C
1 Aircraft Assigned
i ' 1 i Speed (knots) . Altitudes (feet) !
1 F-4C
_„ 1 50C-9CO 550 j 0-^)0
420 500-900 l
0-400 j
360 0-400
A-6A 36C 500-900
o-4oo j
F-105D 420 400-900
Table 2-1 Aircraft Assigned Speeds and Altitudes
0
i
2-9
- •-' ^■■.-:..^—-^ - ^g
-
SRWHPSÄiPPPffWW' —,-Ä—^iw
V-r-ip.,^^,>'^^. ■:*;;: ■V>u^)yw^*^-,^K^:^'/'?fl^-M^^ ■
«»»»«wwwww-v«« ^"•»wmtf-w.m*""": ■ ■
-
Ilf^ncpipiiiiiijiiiiiyiiu,,.! wtfrmmmmm ^lyffwswpqwuffyoff
a detect event was recorded for an observer after he made an
Inraage
eBtimate, the detect event vas discarded.
3. DMCA REDUCTION.
a. The ground observer event data were time-related to Teat
k.l
aircraft position data by use of computer programs. At this
time, the
performance measures were computed, as were the values of the
Independent
variables to be used in subsequent statistical analyses.
i !
; I ;
b. The Independent variables calculated for each event
included
such factors as the aircraft slant range (SR), the sun angle
(&A), and aircraft angular velocity.
c. The performance measures Included detection distance,
algebraic
range estimation errors, and the pilot acquire minus observer
detection
time interval.
k. ASALYS1S PIAN. This section describes the general approach
used
for statistical analyses of the data. A more detailed discussion
of
the analysis approach and methods Is presented in Appendix
D.
a. Descriptive statistics Included histograms and cumulative
frequency plots vhich presented the frequency of occurrence of
the
specific values of the performance measures and independent
variables;
and scatter diagrams (SCD), vhich graphically presented the
concurrent
frequency distributions of two measures, such as the concurrent
(or x, y)
distributions of detection range and aircraft angular velocity.
Also
computed for each variable vere the mean, median, standard
deviation and
maxlmua and minimum values used.
3-2
UMWÜiffW bj^piflWMM iö" w-i"
■ ----^ - ■-— ■- ■ ■-^..-. ■■■.—i^-i-v*.-.-
""'---'M^^ifa^MMIMlJIllMJMIMlL
-
.^rastyv^V-iw.^ piiiiiiMiiii , wmmm : ■■*!*.
I
lM»rt>»'l»»*Ä>W»W.' »iW«>-ft»B«B*iail*
-
^PWWWTWWTWPP.L . IJ11.11!
lWI,W!IJ.ip^'W)1l,1*;:»r'wW'«wT»'-v■»-'■»■>'?■'■-■..^ »rww
I
VO
nn
^
o
CTN
CO
vo
o o
o o
o o
rH O
PO
cv
H O
r-i O
r-i C\j
H CVJ
--1 OJ
r-l CM
OJ O
O O (MO
CO o
C\J o
w o
cu o
CVJ o
CM O
r-i O
H O
rH O
r-l O
ir\ o
-* OJ
OOOJ
CU CV. ^ o
cvj oj
oj cvj
H OJ CM
M O
■H -P C -H H ^ ffl c rs o
o
C3
-P OJ
SB a
VT) I
K
LTN o
-Jt O
lA O
00 ,
-3" CVJ
-tf CM
IA H
J- LTN
lAJt
Jt LA
on LA
5B a
i
on i
O r-i
O r-(
on i
CVJ o
OJ O
onoj
on J-
oj o on on
CVJ O
CVJ O
IA o on o o, o
Cn r-i
o o
o o
o o
o o
O i-l
O .-I
O rH
O H
> a
< \D
I <
CM O
CM O
H O
O O
H O
H O
CM O
CVJ O
CVJ O
CVJ O
is a
o LA O H
I
CM on
IA LA
MD LA
\D LA
MD LA
M) H
MD r-i
IA H
IA H
ON on
CO on
co on
t- on
3= S
o
on o
m o
O o
O o
o o
o o
O CM
O CVJ
O r-i
O H
H CM
O CU
O OJ
H CM
I <
on H HO
OJ o
OJ o
CM O
CVJ o
CU o
CM O
CM O
CM O
CM O
CM O
CU O
CU O
> S
LA on
LA on
LA on
LA on
ON on
ON on
ON j-
o CM I-I
CO CM
O CM
CO CM
Vß CM
VX3 r-l
IACM
LA CM
^ a
o LA O U r-l -J
(*( PH
a *
g
i- V 11
0
0
E
0
■p
U
0
+3
M c C
•H C M H
00
1»^ v_
r-i ^ r-
JH ^ ^^ E-
^ ^
3-^
^iSJBSi««
-
liWBWHW|lll|||WilPHBWipiliWi^
L.J
Ü
II I
ID
..».•»rw.*!tr*«,.,*:.l«',if-W.V('»«,*W.,Jl
tttt^inMwiwwwiWMiMyww
Q rif •
I 1 flight characterlstlcB could have been adjusted by
statlatlcal correlation
techniques. In the absence of unmask data, the variations In
detection
range could not be meaningfully attributed to any of the
Independent
variables. Similar confounding of uncontrolled variables with
test
variables affected the analyses of the distance estimation
errors. The
net result vas a decision to limit the analyses to the gross
descrlptlve statistics for each of the performance measures.
t
I n Hi
3-5
WBigWWWIWWIiMW
...X...t.i.,„.:L..,..,üiAv^.ui.i-jLl aa^di&a a^aüflkia^^
'->...,'.a^-^^'aix.. üiif^i * ^mä - imiiiiiiiriii ■ülilii
-
mmmamaas iPilPgpilliPlllBWpiPBWPPPiwp iMfiivfimmw^-KTT^''
MMtWMMMamiw' vatiriiHiMiattMSH
!
I I
UNCLASSIFIED
I
SECTION 1*
RESULTS AND DISCUSSION
1. (Up^HODUCTIWI. xhiö secuicn preöc^t^ the results and
discussion of the
da^a pertaining to aircrar-: detection, the posttrial
questionnaire, A (time
differences between aircxe* ac^uisiöion and observer detection),
and range
estiaaiion. The major eivphaais in the presentations is on
descriptive
statistics. Results of the more sophisticated analyses are not
presented
for the reasons already cited in Section 3.
2.(CpETi;G'TI0N.
a. The functional relationship between cumulative percent
detection
and aircraft slant range was determined for the 678 observations
available
in the data bank. These data indicate that the aircraft were
detected
50 percent of the time before they were approximately 5000
meters from the
ground target. Ninety percent of the detections occurred at a
range of
3000 meters or greater, and ten percent of the aircraft were
detected at
12,500 meters or greater. The relationship between cumulative
percent
detection and slant range is presented in Figure k-l and may
be
apprcKimated by C = 3-639 -.37098 Ln R, where C = cumulative
percent detection, and R = slant range in meters.
I A' s
I
b. Figure 4-1 presents a comparison of the overall detection
performance
obtained in the Oklahoma/Arkansas Test and the results of
previous detec-
tion tests reported by the Human Engineering Laboratories (see
reference l)
and HumRRO (see reference 5) which were conducted in a desert
environment.
The HumRRO test used a search sector of less than 30 degrees,
and temporal
early warning was provided within ? minutes of the trial. The
Human
Engineering laboratories' (EEL) test used search sectors up to
360 degrees
i
I , ..s.sif1? ^
:!\|?
1^,^,n«JI««w«^««K;*-w**mo*wa,*' •
«^iUÜt^,,.;,^-^ . t&ii id. .^.jLiiiii ■:—'■ i
jajMiiiiiiiiiii^^
-
ijBl^Tl^^TiWBPWyWiriB»^ ; -ly*. :.- ,y~v,T~>W-W, "1
%
: i ' j
\m ASSIFIED
0) •rl •o +:> CO
o
s: ■P ■H
0)
ft B o u c o •H -p Ü 0) -p n R -p c 0) o ^1 0)
I
3 •H
i j
I 1 !
i;
\ i
.,J^-,;;-,,^i.'_;;>iI.;üU:.;i..j^.i.^ ...... ^^^■.,.^^^ MM
MM
-
m* mwiMJUPM.«!..; mvmmmm «F^WWPW gB'iB^ljtjjIfWUf'VliP
WiWWWwuaaiMiwfliwii^wiHMWHBiiiMwiM
^■.■.Lj.-flitmiW;-l»iy»i'.'«fT-.'"i,*
I I I I
0
t f «■
ar.d no temporal eaxly warning. The lest 3.1/3.5 (-i?)
Oklahoma/Arkansas
detection data is contained within the envelope ascribed by
these
earlier tests. Further comparisons are tanacas since different
aircraft
are represented in the thrc- i*, :s. The HurüGO test used F-k,
7-100, and
T-33 aircraft, ^ne E2L tes^ . ...d F-1C0, T-33> and F-86
aircraft. The
different results o'dtained , "-he "chree tests may be, in part,
due to
differences in terrain environ.'sent, search sector used, early
warning
conditions and aircraft characteri-btias. The KumRRO test may
represent
the ideal field detection situation because of the narrow search
sector,
iffiainent early warning, desert visibility and unobstructed
terrain used.
The HEL test may represent the worst caou detection situation
where excel-
lent terrain and meteorolcgical conditions are employed because
of the
large search sectors used and luck of temporal early warning.
The Test
3.1/3.5 (iw?) detection data represent a more typical field
detection
situation where warned observers are deployed at tactical ground
targets,
and the terrain and meteorological environments were
representative of
tactical conditions.
c. Percent detection as a function of slant range does not
reflect
aircraft characteristics. A measure of detection performance
which
reflects variation in aircraft size and heading is the aircraft
subtended
angle (ASA). ASA is the angle subtended by the diameter of a
circle
having an area equal to that presented by an aircraft at a
specified
slant range from the observer.
d. Cumulative percent detections is correlated .965 «ith ASA.
This
relationship can be used to predict the probability of detection
given an
aircraft slant range. The equation describing this relationship
is:
C = 2.717 + -Bio Ln ASA, where Cp = cumulative probability of
detection,
and ASA = aircraft subtended angle in radians. ASA may be
approximated as
M ASA = 2 x tan
J target area
4-3
mm
^st^k^MJaatetataa^^ mimmmm^mmmm^mmmmm
-
l(IlSW»ttW«W».V*K.;-'AVyrt'MW«»V'-'r™;^.#"«-fl!
"irrii.»*»»..',
; i ^;^iHt.ü
e. The area precervced by the target used irt the ASA equation
may be apprcoc-
, where imated as Target Area - d r~2 2 2 A,, x R„ + A. XH +' A^
«VR - R. - H
As = aircraft side area,, .-_ = uircrai't bottom area, A^, =
aircraft frontal
area, R = flight path offset from o";
R = aircraft slanx range of interest.
area, R = flight path offset from observer, K = aircraft
altitude, and
f I i
f. Figure k-2 presents the cumulative percent detection as a
function
of aircraft slant range for each of the four ground targets used
in the
tes-:. A quantitative description of the target sites is not
available;
however, ground survey data was available which described vhe
computed
masking conditions at the target sites. Pig-ore 4-3 presents the
computed
unmask profiles for each site. These profiles indicate the
altitude at
which an aircraft had to fly in order to be visually unmasked
for each
ground target as a function of ground range. These profiles
assume that
the aircraft flew on course, that slant range and ground range
were
equivalent, and that no near mask such as trees obscured the
observer's
vision.
g. Inspection of these profiles indicated that targets Wl and W4
had
very similar unmask profiles. As might be expected under this
circumstance,
the cumulative percent detection plots for targets Wl and Wk are
very
similar. The cumulative detection functions for targets E3 and
E4 appear
q.uite different both from each other and from targets Wl and
W^. On target
E3, the abrupt change in mask altitude at a ground range of 65OO
meters
(Figure 4-3) represents a ridge line. Inspection of topological
maps
indicated that beyond this ridge line, aircraft could fly below
the mask
altitude and would be visually masked. It may be inferred from
the cumula-
tive percent detections (Figure k-2) at target E3 that the
aircraft were,
in fact, rarely available for visual detection prior to crossing
the ridge
line.
I i
k-k
v^^,^.v.J,J,,;^^^i»,&^^.»iatotoi^^».J:^at^.ä^ gÜÜHÜBI mmm
mmimmmä ■m*mi
-
./.cPitilHI^ m i. wummmmmm ^mmmtww.wu '~r ^~-
„.,.. ..«»«HI!S«'»IW**
I I I I
U:ia.A:>S!FIED
lt-5
V •p
■ s
w 1 ■p "« V i to ä >;
EH ^5 ta i o i Vl 1 o
'■^
s Ti XI A c 4. o •H i ■p ,M o ■./>. 4) ■ii p « s i p ,1 £ :;|
o ''4 Ö .1
PL, :* 0> 1 > ;:l •H P s 0) ■s
CM I
•H
'i.■ ^■-■•: .■..!:-.;.,:.-t..^.A*rtJtv -— —■ ■'—^^— a - *mm i —-
--——^ggamiiiaMgg
-
(IWIWW'IWPPI»"1!!""1""'»"!
■r) ^\f\tö
5
8 8 8 0 CO VD 01 rH H
T
\. v
\.
_ h-
_-*
- 0\
>0O
- t-
VD
•C\J
^
o o 8 0 cy
(^SSJB^. aAoqB ^.aajp) apn^j^V MS^W
8 o
[0
m
15
•a § o Ö o
O lH
(0
o
^
(V I i o o Pf 1
!
U-6
w^amiiwiMwwmi»»»»^
-
^WllSIIP^
-
5SHf8R8!B!SBS8B5P!Q!P5!^^ wpw;*p,iiipi' Y7**vi'''*'""''
Table ^-1 Responses to Aircraft Background and Smoke Questions
(u)
4-8
! S
. » *■
,f.:M O-' • 1
Aircraft Type
No. of Trials
i
1 Sxaoke Question Cloud Terrain Total
F-4C 523 £.7.oke 2G&
1 121 30 365
Kc Saoke 9^ 49 15 15S
A-6A loO Smoke 49 14 5 63
Ho Smoke 5h -7 21 92 |
?-105D 66 S.aoke j-X 20 -- 3x 1
No Smoke 23 11 1 35 |
Totals 439 233 72 749 |
r.BS'«al«.8fr„J«-.:«lE'»Ä«ä*~.
-
mm. IRIWpHWPPPi!WB W^Sf&GtKII&lfllflll^^
iiWWiiwM>WW*WWWf .vi-ü^i^wir^iW^
'M f
:-Cw.-.;J. *■.. ,...;.vantages of
.1;. -.v.rre -,. >xx*...a^ely eq.ual.
..■^'.Än Äo-Way
vl.ly reliwJlc (__
-
WProSÄÄOwSlIlWBS TjI'.^'DW TTTKI ,P'P"tJiJr»J--t;-
■ ■■■■■■iiriiir^H' .SSIFIF.n
the aircra-ü has the aiv^aü^ge in that evasive action can be
taken, and
preparations can be xade for the type ox' ground defenses
expected. For
these reasons, Infoxuatioa on who detects first, and by how long
a time,
was considered important.
b. Perforoance Cver All Trials. Figure k-k is a histogram
showing
the distribution of JL across all trials. The overall range of A
is
quite large, ranging fror. - Tx-^v seconds to -.- 79*^5 seconds.
The mean
of +6.51 seconds indicates that, on the average, the observer
detected
before the aircrew acquired. Apprcxiüately 28 percent of the A
are
within the range of — 5 seconds.
t 1 V • *■ 1
:■■
»■ 1
c. Differences Among Targets.
(l) Table k-2 presents the mean acquisition and detecting
ranges
and A data for each of the ground targets. The differences
obtained
between the targets are not difficult to explain in terms of
gross charac-
teristics of the targets. As shown in Figure 4-3, unmask
conditions for
targets Wl and W4 were approximately the same. However, Wl was a
radar site
consisting of vehicles and equipment painted olive drab located
on relatively
low ground, while W4 was a SAM site, containing missiles painted
white, which
was located on the side of a hill. The observers detected the
aircraft at
the radar site before pilot acquisition far more frequently than
they did
at the SAM site containing white missiles. Since the aircraft
unmask con-
ditions were similar, the mean detection ranges for these two
sites differ
very little, while the mean acquisition range at the radar site
was slightly
less than half of that for the SAM site. Although the observers
detected
first only 50 percent of the time at W^, the mean detection
range was some
600 meters larger than the mean acquisition range. This
difference was due
to the fact that there were a number of detections at very long
ranges (over
20,000 meters) at this target, with no aircrew acquisitions
occurring at
comparable ranges.
k-10
•CONriDCNTIAL
*^
-
^SWW^nUWfSIWJBBlBWSrami WW^^^ J 1 1" II NütW* ■•-
vanMtMMHm -s •• w»n '^"ffi
• -^- -^....-n, iiuaüMmi
-
WRWRrP mm. mmwrtnn
ii \ •i ^;3IFIED
1 ■large; t,
■ ■ ' , ■
1 l-ilo-o
of : ..JOU-^^:1C:. C0üerVa"UiO.r,3 ;\;.;.0e
I /', ,.. „,.■*
0"b.,Qrver Mt an
Detection ;-Sa,Ilge
(itete rs)
Percentage of
Positive
4
Mean
4 | (seconds) 1
Wl 1 I
l63 , 24H-0 1 6091 79 16.96
W4 1
ITT j ^139 1 ^T31 50 2.85
i
-3
... i
i2T | i5]+i 31+94 55 -0.3T
Si. I4O 6336 ) 7694 56 5.20 j
607 4| CL 6231 60 6.51 j
rable k~2 Pilot Acquisition and Observer Detection Statistics by
Targets (u)
4-12
ymcvA^ •^m^MMif^i^mi-imm'l'm',m'Mmm
.ü:.;,:>,;.;i,:;^-';::„..;:.:;:.;f.;
MmMmMääSläMiämä^Mä^t^^lMiSill^ilik
-
55HB55)SpwKra!Be55siSISBKn«s^^ aWS^Wf^»*»*"™ . ! ..
lAmwwrT,'.-:.,-r,^,.m-,..
■.■»m »JW«■»IWI'f'WMW
ÖNCI./I^^FlFi n
I
it
i I I
(2) At target E3 the acquisition and detection ranges were
approximately the same. This target was an airstrip so located
that unmask
was normally relatively short, occurring as the aircraft flew
over a ridge
line approximately 6,500 meters from the target. At this unmask
range,
aircraft should he plainly visible. Also, at this range, a
target as
large as an airstrip shoulu have heen plainly visible from the
air. There-
fore, it is not surprising that pilot acquisition and observer
detection
occurred at approximately the sane time and same range. While
the observers
tended to detect before acquisition somewhat over half the time,
neither
the difference between the acquisition and detection ranges nor
the mean
A for the target indicate any advantage for either aircrew or
observer.
(3) Target E4 was a bridge, and unmask range was quite
large,
although rernask was easily possible. Both aircrew acquisition
and obser-
ver detection ranges were larger for this target than any other.
The
average A data indicated some advantage for the ground
observer.
d. Trial Effects.
(l) For any given trial, either none, one, two, three or all
four
observers stationed at a target could detect before the aircrew
acquired.
Table 4-3 presents the percentages each of these events occurred
during
the 134 trials where responses were available for all four
observers. For
k2 percent of the trials, all. k observers detected prior to
aircrew acquisi-
tion, and for 23 percent of the trials the aircrew acquired the
target before
any of the observers detected the aircraft. The figures in the
row labeled
Expected Percentage were obtained from the Binomial Theorem by
assuming that
the probability any given observer would detect before aircrew
acquisition on
any trial was 0.60 (that is, the same as the proportion of
positive A that
occurred). The analysis indicated that either none of the
observers or all
four of the observers detected first far more frequently than
would be
expected. (This indicates that the observations were not
independent,
as assumed.) Although this result may be due to differences in
the manner in
if-13
UNCiASsini n ttSS*ß»«*««!««w>*«
,w
"-■■•'—•■'■-*'-■'- —■'■ -^A-- ■■ - - n ^m^^imjgiii
-
n'rWWHPWV^'1^
i;:;CUs;
4
Cbserver Detection Before Pilot Acquisition
K ■3 2 1 0
j Percentage of Trials the Event Occurred K2 lv 11 10 23 j
Expected Percentage 13 35 35 15 3 |
Table 4-3 Percent of Observers Who Detected the Aircraft Before
Pilot Acquisition for Trials Where All Four Observer Responses
Were Availaole (u)
k~lk
„oL!^^
L-i.i:,i&ijiA
-
!"' wmmarzm
i
I I I I I I I I I I
.u1 .»■'i'"'rtA't fffl1
-
^göiWRpronwgj^-rtfawtt^Ju^ '■"-7 i,IW.^..
fW.n.PüwupiTTViT-i."
■' t /•' • ( v^m
no cleLi. asensus was reflected in the observers' responses. The
obser-
vers dia ctgree on the following points:
(l) smoke output caused the F-kC to be the easiest aircraft
to
detect.,
(2) each observer tended to return to his "favorite" lookout
point at each of the targets,
(3) early warning information was heard via the radio located
at
the target area during the no early warning test trials. This
information
was useful for target Wl, but of little assistance on the other
three targets, and
(k) the observers also reported that they had a pretty accurate
idea
of where the target would appear by the end of the first test
week. This con-
tention is supported by the size of the search areas reported
for the four
targets. On the average, the search sectors were reported to
vary between
approximately 50 and rj0 degrees even though the observers had
been instructed
to search a l80-degree sector.
6. (C)RANGE ESTIMATION ACCURACY.
i .
1 I
11
a. The range estimation results are presented graphically in
Figures
4-5 through ^12 and are summarised in Table h-k. These figures
and table
describe range estimation performance under the following
conditions:
(1) test conditions (before training, after training, first
week
field test, and last week field test),
(2) direction of aircraft flight (incoming or outgoing), and
(3) engagement range to be estimated (400, 800, 1,500, and
2,500
meters).
k-16
^^».i-.wi^i^^i,,^ iüjüüÄ ^jtajbjaMiteaijaiii^^
/ i
[I
Li
-
i i.Jifl'i^.i. f. ■J*'.L'ri^F7;ilUlfl?r'A7iJinWP.J]!T.1 ?i^"'2^
WBWPWIJI,,W,Trp' "•■IpSIfJlp:" i
f,.Y,NTnVl
-
j
-
„.., , „.,,-,,,■»■';..W■*l««|W,fl','l"'
I
I I I I I !
I I I I I I I I
' * * i ■
Figure k-o (U)
I+-19
4
.^V^^V^i-iiiLAiV^Vl^t^i^.l^^
-
WBWIWHWlgWIWIW »wwipimw 'M1'.1" "iw 'n11
;L/wolrlLD (
' ?
■'aASSlRED
Figure k-J
k-20
(U)
ttfwataiWiinuwHtMMnu'
-
^.3mi^^mr^^v'i^m,Hmm^m^ym^f' f^WfFT'ÜK'ffWVH"1 VpiW-W '■ ■-"
w.™^*^,,,,
"MWttff.JÄWtMM't ,'ftf'r'«»-^"^w* ■i,f.
^ I;
I
I I i i i I
I I
^infllOMPIDENTMkL
Figure k-1 (u)
k-21
i,Mn A^IF! CQNPIOEWTWE
-
. i .r^wWÄ^Ä-rw^'^u 5--t^>«^ 5SÜ5 fpnmwHmW^illlpnpKHRn
P^np^PIBP^PiP WWPiPS?^
UMC1 ^IFIFD
•I t
( ('
a> o
i ^
-OL.
.6
SS 1^ » o
Figure k-7 (U) K-22
COWFIDEIHf^b
8 CVI
o O O
er. c o CD
•4-> 3 O
o o in
O) c (0
■o
c O)
•r" (/) CO <
icn
■'v.i.'.^■MI-.VK ■*-V»i..;'ivu«^.i'iK-it.^T..
^..:.,.^.^.WJ.^. .^.^i^^Wk^vi^llH HMukitt
-
"WÄJ'RWWl'ISIW
•^•m.itox-r.uvwK-vmf^MKif'-:-*!
1
» K. 4- i
i |; 'S
^ f 1J»W•»-^S'■'f*^llW*,M*"™,',
c *: n ^ ^ ^
*3 cc ^ ^ ^
-
PBfflqwQwn 18! wsmmamm ifm^rmm*
I Mirxj ^oo
4IONI IDENI Mi.
i f r !
o
SO V? en
a, +1 g» II S
I n 2
c
o
CD
D>
c CC
< -I
ig Q
I «I o
o o o
o o ro
O to
o o o CVJ
o o
en c o cr. +J 3 o o o CM
OJ CD C ro a: X) ai c CD
■r- CO CO < I
Figure k-Q (u)
n^Ci ^SIHED
-
y$!$K^&£'i'7^vZ%!2i!l RPPMü iqmvmm
Wf1tPB»)W^^5«iWW^.Mif(^'B'ii*E,'n*iww»i
TmpW f-J'.ll1«i|i"iMWWvMr«'T:",-Tu-r-.™.iTn-..T
,.,.„. «™m^.»4w^*Mi^*WWv6tj«J*M[Bff'f*,W
I I I I
I I I
I, lASSIFIE
Course Assigned
Range
Mean I and Standard Deviation
Before Training
After Training
First Week of
Test
Last Week of
Test
Inconiing
400 X 797 357 6UQ j 83O 1
-
»JK'ÄSWJi*»:
b. An inspection of the before and after training data
indicates
that, in general, training served to decrease estimation error.
Negative
training effects, which were partly caused by an instrumentation
bias in
the training procedure, were obtained for some of the longer
ranges on
the outgoing portion of the flight path. The training was most
effective
in reducing trial-to-trial variability of estimation errors, as
shown by
a reduction in the size of the standard deviation for each
assigned range.
In general, the training levels achieved by these observers were
comparable
to the levels of performance obtained during previous range
estimation
training investigations (see reference 6).
c. One objective of the range estimation analysis was to
compare
proficiency during training, which was given under one set of
conditions,
with the test performance of the trained observers obtained
under a com-
pletely different set of conditions. The training was
accomplished in a
desert environment under excellent visibility conditions with
one aircraft
which flew a constant speed at two programmed altitudes and
offsets. The
testing, however, took place in a semi-mountainous region with
very high
humidity, which tended to reduce visibility. The test
environment also
included three different aircraft which flew numerous speeds,
altitudes,
and offsets over four different test sites.
d. It was expected that the influence of changing environmental
and
stimulus conditions from training to testing could best be
determined by
comparing the end-of-training scores with the scores for the
first week
of the test.
e. Figures 4-5 through k-12 indicate that estimation errors
increased
from the training condition to the test condition for the
1500-meter and
2500-meter incoming estimates and the 80ö-meter and 2500-meter
outgoing
estimates. In general, the average estimation errors slightly
increased
from the training to the test environment, but the variability
of the
estimation errors drastically increased between the after
training (T)
\\\CV r ^ c^Ut
k-26
*,M*M*j;iwmit^tmi!i^t^M*w''*^'w''™'~''*'~~'''
■ -^-**tLi*-'L-^:a,:li.\.^A..tJ-£. .^.^i..,^^^»^^:^*:^,^
-
iPPyippimW*»WBW*piWlW^ 'W7,-^rm5"mnw»w»'w^i*r!T^-i—»■"^" ■
-VM^f^TW;!«,'!''.'*!' ■;^t»»«ll«'i»f01?*-vf!(»'
i 1 l
1 I i I'
l
l
l |
|
|
WMW
and the field test first week (P..) testing conditions. Voss and
VTickens
(see reference 7) found that when observers were trained to
estimate a
range of 1,7°° yards under one set of conditions, the accuracy
and variability
of estimation remained approximately the same over a 6-day
period. Since
there was a 5-day period between end-of-training and testing for
the
Oklahoma/Arkansas test, the conplete change of environmental and
stimulus
conditions appears to be at least partly responsible for the
large increase
in variability of estimation.
f. At the end of the fifth week in the test environment, the
overall
variability of estimation remained approximately as it was after
the first
week, but the accuracy of estimation was decreased. This finding
is
inconsistent with results obtained by Horowitz and Kappauf (see
reference 8).
They found that range estimation performance after training was
stable for a
period of 60 days without additional training.
g. The most consistent result of the range estimation
evaluation
was the occurrence of large variability in the estimation errors
during
the Oklahoma/Arkansas testing. As shown in Table k-k, the
standard devia-
tions of the errors were very large. These results suggest that
retention
of this skill deteriorates rapidly over time, particularly when
no feedback
concerning error magnitude is available to the observers, and
the environ-
ment is much different from that used in training.
h. HumRRO has reported a series of studies concerning range
estima-
tion accuracy (see reference 6), which included a conparison of
the
accuracy of judging a 350-meter distance with and without the
use of an
occluding or stadimetric aid.
i. In one of the HumRRC studies, men were trained to estimate
350
meters distance to an aircraft. One group of men were trained
using
techniques similar to those subsequently used for the 3'l/3'5
observers.
lf-27
■■■ • ■•, 00
*jLlui1,'u,-jLW*!*iJf„i£.*-l „f:^: L , ^M.mri.iaw.mmfnii^
■il^iä.^£LmA^iÄ.i£ti!. mmmmmmmm
-
BP^BgBffKWBWWiwcapiasmsgp^ wmiv* K"
'•vvl\r']'m^W>MMr^WK^vr^^Tyr>^rr'f'^-'nr'^-rr^- r ""*•
i-w-?
ira ^\f^D
A second group of men were trained to use their index finger,
which occludes
a liaison aircraft at approximately 350 meters, as a stadimetric
aid. Use
of the occluding device was found to reduce average bias and
variability
in comparison with the unaided training. It was incidentally
learned that
the front sight guards, or tangs, of military rifles also could
serve as
the job aid for determining when to open and cease fire against
aircraft.
i. In 1968, HumRRO began a study to identify existing ccmpcnents
or
appendages on US air defense weapons which would function as
stadimetric
aids. The results of the 3.1/3.5 distance estimation tests
support the
need for some type of simple job aid which gunners could use to
estimate
the open and cease fire events.
%
^-28
ynCU\SSlF
Saaaa^iütAtaakiiiitefaaa^^^ - - ' 1am „it v.-l ■ j.
V-WsirfLdi'-'Si..,
-
.waWmP^ittW^W^^^lfthlAWW^'^W»^^
WPPWWPPffüPWffWW ■ -DU-J'.^i PBJPffWgljilP^g^
BMMWBWimBMflWW^l^
I Li
1
APEENDDC A
BIBLIOGRAPHY
The references belov are listed in the order of their citation
in the
body of the report.
1. Wokoun, William. Detection of Randan Low Altitude Jet
Aircraft by-
Ground Observers, Technical Memorandum "J-So, US Army Ordnance
Human
Engineering Laboratories, Aberdeen Proving Ground, Maryland,
June i960.
2. Zimmer, W. D. and McGinnis, C. F. Redeye Target Detection
Study,
Technical Memorandum 1072, Army Missile Test and Evaluation
Direc-
torate, White Sands Missile Range, New Mexico, March 1963.
3. Fimpee, M. D. "JTWRAP," Sandia Corporation, Division 9^27,
Sandia
Base, Albuquerque, New Mexico, undated.
t i
\
10
k. Clyde, D. J., Cramer, E. M. and Sherin, R. J. Multivariate
Statistical
Programs, University of Miami, Biometrie Laboratory, Coral
Gables,
Florida, undated.
5- Wright, A. D. The Performance of Ground Observers in
Detecting,
Recognizing, and Estimating Range to Low Altitude Aircraft,
Technical
Report 66-19, Human Resources Research Office, Alexandria,
Virginia,
December 1966.
6. McCluskey, M. R., Wright, A. D. and Frederickson, E. W.
Studies on
Training Ground Observers to Estimate Range to Aerial Targets,
Technical
Report 68-5, Human Resources Research Office, Alexandria,
Virginia
May 1968.
I i
A-l
fiBtfeiW^WWW***"AiW
j^adteaaU^ak^^
-
WMiJ»ll.ini>,J.yiiiww|jiWiwi«*«M-i.?w-i HBPWWF ICfWpWffHT
7. VusS, N. A. and Wickens, D. D. A Comparison of Free and
Stadlometrlc
Estimation of Opening Range, Memo 29, Applied Psychology Panel,
Project
N-.1C5, K6]11|, Office of Scientific Research and Development,
Washington,
D. C, 19^.
8. Horowitz,,, M. W. and Kappauf, W. E. Aerial Target Range
Estimation,
Report No. k, Project N-lll, National Defence Research
Committee,
Washington, D. C, 19^5-
t
A-2
-
Tmmammmm mmmmt*
^WIIIlWlill'y'B^IIIWIVWWIIWWlW^^ ■■'■M',if"r-,-7[r-ir'-tfr-TTin
imrn WIT IHI
mwmoariimm'imMniiiitmjiimmWfiiiwiatrfflmwiftwrwwfWiOTtriiwnMW^
|1"'
t-
I I
0 0 D 0 Q
Q n
i n I 11
APPENDIX B
RANGE ESTimTION TRAINING
1. INTRODUCTION.
a. Previous field studies conducted at Fort Bliss, Texas,
concerning
range estimation training methods indicated that instruction
using immediate
knowledge of results is the most effective and efficient method.
The use of
this method had resulted in rapid improvement in a short period
of time with
smaller errors than when other methods were used. For these
reasons, the
method of immediate knowledge of results was selected for
training the l6
ds servers.
b. The training was conducted over a three-day period. At
the
end of each day's training session, a test was given to
determine each
individual's status as training progressed. A test was also
given before
the first training session to provide a performance baseline in
order to
evaluate the effects of the training.
2. DESCRIPTION OF TRAINING ENVIRONMENT.
a. The training was conducted during late morning and early
afternoon hours at Hueco Range No. 2, Fort Bliss, Texas. The
relatively
flat desert terrain provided for meteorological range of
approximately 75
miles. To the near west and distant north there was a
mountainous back-
ground, and northeaut, east, and south there was sky
background.
b. In order to reduce the possible Influence of terrain
features
as cues for range estimation, three training sites, several
thousand
meters apart, were used. Training was conducted at a different
site each
day.
B-l
I '
■ ..■■ ^JMJrfUMaaaaJa.ttaihM^ sfj M jcBSSifiMfefl i i am
-
^Wt^PWBWCTBEiW-^-M^Bs^Tnry^Wy^CTWTrWffSff^PffgiCT
mUHWtWBWiiipwiiwjpin yipiPMi, >i
c. Two parallel flight paths were set up for the F-100
target
aircraft to follow. A "red" course passed 200 meters to the west
of each
site, and a "yellow" course passed directly through each site.
The air-
craft flew at one speed, 400 knots true air speed, but used two
altitudes
in order to vary the aircraft aspect. These altitudes were 250
feet and
750 feet. Alternately the aircraft flew from the north and
south.
3. TRAINIMG PROCEDURE.
a. The observers were initially instructed as to the nature
of
the training. They were then positioned so that each could see
the entire
flight path in both directions. The instructor prepared the
observers with
a warning (REAUy) a few seconds before each signal to estimate
was given.
As the aircraft flew over the course, the observers made two
estimates of
the slant range to the aircraft when a signal (ESTIMATE NOW) was
given
by the Instructor. These estimates, one while the aircraft
approached and
one after the aircraft passed over the site, were recorded by
each obser-
ver on special record forms. Immediately after the second
estimates were
made, the Instructor announced the correct ranges at the time
the signals
were given. At that time, by referring to his record form, each
observer
could immediately determine his error of estimation.
t *
b. The observers had been told specifically that they were
going
to be trained to accurately estimate four different ranges;
1+00, 800, 1,500,
and 2,500 meters; but that during training, the signal to
estimate would be
when the aircraft was anywhere from 300 meters to 2,900 meters
from them.
c. During each day's training session the aircraft flew 36
passes,
18 in each direction. On each trial (aircraft pass) two
estimates were
made, one Incoming and one outgoing. Over the three days of
training, each
observer made a total of 2l6 estimates.
B-2
•-•—-- ■■ -- ci^te^|Mi
-
snsssnm RpmniPHMniifni ippniMp 1,11'mi . mpa Ptin
i"/f'-»""^*l|l|,| »J iH»fT'«"'P ■
i.-w»iwuw(«y.*115»'.)r-':;" •'
•n^^v^v^-A'tV'ff'^w/fJi/wi^frjwWMmMJ^^
Tr^n(^Mi*«:«iW»rtwtw,/iwrf»»i,'^,H"i»ii^BBiifflm»-Wm
H'j»wjimePtwln«««*W)r,#i« ^iW«M»»K.»«WW«*««w*ll1WBiW/**5iP^fl(
I
0 n D
u
! l-i
0 i ID
u
d. The signals to estimate given by the instructor were
based
upon a timing system which relied for accuracy upon the ability
of the
aircraft to maintain a programned speed and course within some
limits.
In order to check the accuracy of this system of determining
true air-
craft range, as the aircraft passed over the training site
during test
trials, a "crossover" mark was put onto the event recorder used
to record
observer test responses. By comparing the timing system with
these
crossover marks, it was discovered that errors existed in the
system.
These errors resulted in training the observers to use a
different
"yardstick" than was programmed. That is, when the observers
were told
during training that the aircraft was KOO meters away, incoming,
It was
actually closer to 525 meters away. Table Bl-1 shows the actual
ranges
of the aircraft when the observers were told it was at the
programmed
range.
k. TEST PROCEDURE.
a. A total of four 12-trial tests were scheduled, one before
training ccmmenced on the first test day and one at the end of
each of the
three days of training.
b. Just before the aircraft began a pass, the observers were
told
the two specific ranges they were to estimate on that pass, one
incoming
and one outgoing. They were told to indicate when they believed
that the
aircraft was at the specified ranges. Each observer was provided
with a
pushbutton connected to a channel of an event recorder. When the
observer
through the aircraft was at the specified ranges, he pressed his
pushbutton.
A mark was made on the event recorder when the aircraft was at
the programmed
specific ranges for the purpose of checking each observer's test
responses.
B-3
i . .-■ ■ i-^^M^miM*m*mMk
-
Table Bl-1 Actual Target Ranges During Training Compared to
Programmed Ranges
| 1
Course Assigned Range Actual Range |
Incoming
400 525 1
800 925 1
. 1500 1^5 |
2500 21+10
Outgoing
koo 360
800 585
1500 II85
2500 2C-V0
B~k
1 i
aiA^ii
-
w^m WHW*'^^*'**" ■ HM 'V^V'T« iwmupjfjf ■ v**-""
,. «..«.'0W^;-»',',r*1lt'"»''■■■.•-^": ifcWww^#\iJrnO(.'K^»w
,^/iV»^rv»j»^*-^Möf'i^^ r.^vAT^^wW^'-^^m^SWWSri^^
! il
I I c, The aircraft scheduled to fly the test trials on the
final day-
had mechanical problems and did not fly. As a result, the final
test
trials had to he cancelled. The result of the training has been
evaluated
based upon performance on the test trials at the end of the
second day of
training.
5. RESULTS OF PRETRAINING TEST.
a. The 16 observers that »ere trained varied in age f~oni 18
to
24 years. All had 20/20 vision, uncorrected or corrected. GT
scores
ranged from 83 to 130, with a mean of 111 and standard deviation
of 11.
I;
I n ! U
b. At the beginning of the first day of training, the
observers
were told that they were to be given some training in range
estimation to
aerial targets, but that first they would be tested to see how
well they
could estimate various ranges before training. The results of
the pre-
training test are indicated in Table Bl-2.
c. The incoming ranges of 400 meters and BOO meters and
outgoing
ranges of kOO meters were greatly overestimated. This shows that
the
observers believed that these ranges were much greater distances
than they
actually were. The remaining means of estimates were accurate,
but
variation was relatively large in all cases.
10
0 B-5
-
mmmm ^^^^ü^^^TTT. snsssra rtsw..ü..v.^-.^^i^«s*-^^
ffw^^^yywwwPB'yyc^"";,*yT,if;'^"*»pT^^»aw,^j"'^''»
M '
j Course Estimated Range Actual Range
Standard | Deviation
j Incoming
koo 797 162
300 139^ 259 j
1500 1752 238
2500 2790 282
Outgoing
J+oo 711 15^
800 764 84
1500 1514 123 j
2500 2317 203 I
Table Bl-2 Means of Range Estimates (in meters) Prior to
Training
B-6
■;.jjfci««iW«#«,
-
wemn^wmff^wmm SBBPSPP mu^mi^?7s.zm'*^^*?*^'~ ■••s
t
I I
ti
i 1" i.
APPENDIX C
ARKANSAS POSTTEST INTERVIEWS
l.(U) OBJECTIVE.
a. A set of questions was asked of the l6 military personnel
who
served as observers for the human factors studies conducted in
Oklahoma/
Arkansas as part of JTF-2 Test 3.1/3.5 (NP). The questions were
presented
during individual interviews conducted during the last three
days of the
testing. All interviews were conducted by one human factors
scientist.
b. The interviews were prefaced with the following
introductory
statement:
"Now that we are Hearing the end of our tests here in
Arkansas, we have a number of questions to ask you concerning
your test
activities during the past few weeks. It is hoped that your
answers to
these questions will help us clear up the minor confusions and
uncertainties
that always appear after field testing is done."
2.(C)INTERVIEW QUESTIONS. The below 12 questions were asked of
each
observer participating in the test. For each question the
responses are
indicated in summary form.
a. (u) Qustion No. 1. "We understand that many of the observers
have
missed detecting sane of the aircraft for one reason or another.
About what
percentage of the aircraft do you think you missed? Or, in other
words, out
of every ten flights, how many did you not